Age-related macular degeneration (AMD) is a chronic, progressive disease that results in a loss of central vision and significant functional impairment. It is the leading cause of blindness in Western developed countries 1 2 . Neovascular AMD (NV-AMD) represents 10 to 15% of all AMD cases but accounts for 90% of AMD-related severe vision loss 3 . Choroidal neovascularization (CNV) causes disruption of the structure and function of the retinal pigment epithelium and the retinal photoreceptors. Prevalence of late forms of AMD (defined as the presence of NV-AMD or geographic atrophy 4 ) increases exponentially with age 5 .

Rapid vision loss is a key characteristic of NV-AMD, such that the proportion of untreated patients who develop severe vision loss (> 6 lines) can reach up to 42% in 3 years of follow-up 6 . Additionally, patients with CNV in one eye have an estimated 43% probability of progression to NV-AMD in the fellow eye within 5 years 7 . Patients with visual impairment from AMD lose independence, suffer from fall-related injuries, experience high levels of depression and anxiety, develop annoying visual hallucinations, and require low vision aids 8 9 10 11 12 . Direct medical and non-medical costs and the cost of progression to blindness all contribute to the economic burden of AMD both to the patient and to society. Estimated annual societal costs of all NV-AMD patients in Canada, France, Germany, Spain, and the United Kingdom are substantial, ranging from €671 to €3278 million 13 .

Studies show that the early detection and treatment of AMD can delay vision loss and blindness and thus significantly reduce the economic burden of the disease 14 15 . The primary purpose of early treatment ideally would be to defer progression or to promote visual improvement. Prior to the development of anti-vascular endothelial growth factor (VEGF) inhibitors, NV-AMD patients were treated with laser photocoagulation and photodynamic therapy (PDT) with verteporfin (Visudyne^®). PDT had limited use as it was only approved to treat the predominantly classic type lesion, representing approximately 20% of NV-AMD patients 16 and it merely slowed vision loss 17 .

The quest for alternative treatment options in NV-AMD has been precipitated by the increasing prevalence of the disease and by the associated side effects and unsatisfactory outcomes with approved therapies. Anti-VEGF treatments, such as pegaptanib sodium (Macugen^®), ranibizumab (Lucentis^®) and bevacizumab (Avastin^®), are the first pharmacological treatments to address an underlying pathological factor of the CNV of NV-AMD and to address disease progression without healthy ocular tissue destruction. Pegaptanib sodium, a selective anti-VEGF therapy approved for the treatment of all subtypes of NV-AMD, was introduced into the US market in January 2005. Results from phase II/III pivotal clinical trials 18 19 showed that pegaptanib was effective in patients with subfoveal NV-AMD regardless of lesion subtype (i.e. predominantly classic, minimally classic, or occult). Approximately 70% of patients treated with 0.3 mg pegaptanib had stabilised or improved vision (lost < 15 letters [< 3 lines] compared to baseline) at 54 weeks compared to 55% of patients receiving standard-of-care treatment. Additionally, pegaptanib was well tolerated, with the majority of adverse events being ocular in nature and transient.

Currently available efficacy and safety data for pegaptanib are from clinical trials, which may not accurately reflect pegaptanib's real-world use and potential visual outcomes. Further, there is a need to understand which disease characteristics define earlier lesions in order to identify patients who may have a better response to anti-VEFG therapy (i.e. pegaptanib). We performed a retrospective chart review study in newly diagnosed NV-AMD patients initially treated with 0.3 mg pegaptanib in the US to evaluate actual clinical experience with intravitreal pegaptanib monotherapy and to explore the characteristics of lesions in patients in whom a better response to pegaptanib monotherapy was observed.

The efficacy-related findings of this real-world observational study are similar to those previously reported in pegaptanib clinical trials. In the current study, the mean change in VA from baseline to month 6 was a decline of -0.68 lines (0.07 logMAR), and 70% of subjects lost fewer than three lines. The VISION trials 18 showed a decline of approximately 8 letters, and 70% of subjects lost fewer than three lines of VA after treatment over a period of 54 weeks.

There is no universally accepted definition of early CNV secondary to AMD. The early lesion definitions used in this analysis were based on those used in the exploratory analysis of the VISION study results 20 , matching as closely as possible those definitions using the data we had available. Subjects in our study who met early definition No.2 had significantly greater improvement in VA from baseline to month 6 compared to the decline in VA in those who did not meet the definition (-0.05 versus 0.13 logMAR [0.46 versus -1.29 lines], respectively; p = 0.0235). This result was likely driven by the high proportion of patients with occult lesions in the study while few subjects with predominantly classic or minimally classic lesions were included. This suggests that clinicians when selecting out earlier lesions are defining these lesions with occult characteristics. Our current study found that, on average, subjects with occult lesions treated with pegaptanib had an improvement in VA from baseline to month 6 while subjects with predominantly classic or minimally classic lesions had a decline in VA (-0.01 versus 0.27 versus 0.34 logMAR [0.09 versus -2.64 versus -3.50 lines], respectively; p = 0.0065).

In this study, we observed a significant difference in clinical efficacy in subjects across study sites. In particular, 83% of subjects treated at site No.7 showed either an improvement or maintenance of VA from baseline to month 6, with a mean improvement of -0.04 logMAR (0.50 lines), while subjects from all other sites had a mean decline of 0.11 logMAR (-1.07 lines), with only 49% of subjects showing either an improvement or maintenance of VA (p = 0.0134). The earlier treatment of subjects at site No.7 might have accounted for the better results - the mean time from NV-AMD diagnosis to treatment was 0.3 months for site No.7 compared to 3.1 months for all other sites (p = 0.0808). Whether or not anti-VEGF efficacy can be related to the duration of existing CNV disease has not been clearly defined. In a recent study by Boyer et al. 21 , efficacy of ranibizumab did not appear to differ across the different quartiles of duration of NV-AMD, which is the opposite of what we found in this study with pegaptanib administration.

Although large randomised, prospective studies have confirmed benefit with treatment, the question remains as to which type of CNV lesions responds best to treatment. Different characteristics were identified in previous studies with no consensus. For example, the TAP [Treatment of Age-related Macular Degeneration with Photodynamic Therapy] and VIP [Verteporfin in Photodynamic Therapy] Study Groups 22 concluded that lesion size was a strong predictor of VA outcome based upon observed PDT treatment and outcome while the VISION trials 20 found that anti-VEGF treatment of early lesions (defined on the basis of the combined characteristics of size, lipid, scarring and time of presentation) had a positive impact on VA outcome. A different conclusion was drawn from the MARINA study 21 subanalysis of subjects with minimally classic and occult lesions treated with ranibizumab in which neither lesion size nor duration of NV-AMD was found to have a direct relationship with VA outcomes. Our study also attempted to evaluate which of the subjects' baseline angiographic characteristics had an impact on VA outcome and found that those who met the early disease definition using the criteria of occult with no classic CNV, absence of lipid, and better VA at baseline in the fellow eye had better VA outcomes. There is no agreed upon and accepted early lesion definition at the present time.

This study intended to evaluate newly diagnosed NV-AMD patients treated with pegaptanib monotherapy. Even though we did not provide participating clinicians with a definition of 'newly diagnosed,' we did not find differences in angiographic characteristics across patients' time since diagnosis. However, we did observe a relationship between time since diagnosis and primary study outcomes of VA change from baseline to month 6 with shorter time being associated with better outcomes. The finding is consistent with the oncology literature that shows that younger tumours that are earlier in their angiogenic process of new blood vessel development appear to be more susceptible to certain types of anti-VEGF therapy 23 24 .

There are several limitations to this study. The study included only those practices and patients who were willing to participate. Two sites had a total of four patients who met study inclusion criteria and who were willing to participate, with one patient having maintained VA and the other three having lost more than three lines from baseline to month 6. We cannot be certain that these four individuals are representative of the universe of NV-AMD patients treated in these two practices. Conversely, one does not know how representative patients from site No.7 are to the NV-AMD population either. The sample size of the study limited our ability to perform multivariate analyses that might better support interpreting study findings. In addition, although the proportion of patients losing VA from baseline to month 6 appears to correlate with the presence of classic lesions, this is most likely biased by the fact that our sample included few subjects with classic lesions. Further, the study did not collect optical coherence tomography data for each visit, which might have provided a more accurate summary of NV-AMD characteristics, enabling us to evaluate the relationship between VA change and disease characteristics. Finally, the data were collected for 6 months only; it is not known if the results would have been different over a longer treatment period.

The efficacy of pegaptanib in community-based practices appears to confirm findings from the VISION trials and a published analysis of patients with earlier disease. There appears to be a trend for patients with earlier lesions to respond more favourably to pegaptanib. Due to our small sample size, there was significant variability of outcomes by site and by patient. Still, shorter time from NV-AMD diagnosis to initiation of pegaptanib treatment appears to be associated with better treatment outcomes and enhanced clinical benefits. It is a common theme across the medical and scientific literature that earlier intervention, prior to permanent damage, is more likely to achieve a beneficial outcome. Other large NV-AMD clinical studies 21 22 have attempted to define and interpret outcome based on lesion characteristics. In this study, though, other disease characteristics did not seem useful in identifying a priori responders to treatment. Further research is warranted to fully understand and determine NV-AMD disease characteristics that help predict outcomes. Their results will enable NV-AMD therapy to be targeted to provide the greatest benefit to both patients and society.

Pamela A. Weber was compensated for her work as an investigator in the study by Pfizer Inc and (OSI) Eyetech, Inc. Barbara M. Wirostko and Gergana Zlateva are employees of Pfizer Inc and own Pfizer Inc stock and stock options. Xiao Xu and Thomas F. Goss were employees (at the time the study was conducted) of Covance Market Access Services, Inc., the company that held the contract for conducting the study. The research was funded by Pfizer Inc and (OSI) Eyetech, Inc.

GZ, BW, XX and TG contributed in the study design, data analysis, data interpretation and writing of the manuscript. PW was a primary study investigator, collected data, contributed to the data interpretation and writing of the manuscript. XX and TG carried out the statistical analysis.